Token导航 LogoToken导航

Agent Skill · 研究检索

codeql(codeql 搜索)

用于围绕 GitHub 仓库、Issue、Pull Request、分支、提交和代码协作流程提供辅助能力。它适合让 Agent 查询项目状态、整理变更、辅助创建或检查协作事项,并把仓库中的信息转成可执行的下一步。使用时需要区分只读查询和写入操作;涉及创建 PR、修改 Issue、推送分支或访问私有仓库时,应确认 token 权限、目标仓库范围和用户授权。

codeql 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为接触敏感数据;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 4,224。

总安装 34,608周安装 1,403GitHub Stars 31,689下载 10,864核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:codeql(codeql 搜索)
来源仓库:https://github.com/github/awesome-copilot
仓库路径:skills/codeql
安装命令:
npx skills add https://github.com/github/awesome-copilot --skill codeql
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/github/awesome-copilot --skill codeql

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

集成 CodeQL 搜索能力,支持代码安全漏洞与模式识别分析。适用于代码审计、安全隐患排查与静态分析任务。通过 npx skills add 从 awesome-copilot 仓库安装,需配置 CodeQL 环境。运行查询前应验证数据库连接与访问权限合法性。

codeql 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/codeql

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/github/awesome-copilot --skill codeql;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

codeql 怎么安装?

在终端执行 npx skills add https://github.com/github/awesome-copilot --skill codeql。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

codeql 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

codeql 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

codeql 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

codeql 用的人多吗?

公开数据是累计安装 34,608 次,近一周安装 1,403 次,下载 10,864 次,GitHub 31,689 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

codeql 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。
  • 来源仓库有 31,689 星,维护活跃度相对可查。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
codeql

codeql 搜索 · 本页

34,60831,689接触敏感数据未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

CodeQL Code Scanning

This skill provides procedural guidance for configuring and running CodeQL code scanning — both through GitHub Actions workflows and the standalone CodeQL CLI.

When to Use This Skill

Use this skill when the request involves:

  • Creating or customizing a codeql.yml GitHub Actions workflow
  • Choosing between default setup and advanced setup for code scanning
  • Configuring CodeQL language matrix, build modes, or query suites
  • Running CodeQL CLI locally (codeql database create, database analyze, github upload-results)
  • Understanding or interpreting SARIF output from CodeQL
  • Troubleshooting CodeQL analysis failures (build modes, compiled languages, runner requirements)
  • Setting up CodeQL for monorepos with per-component scanning
  • Configuring dependency caching, custom query packs, or model packs

Supported Languages

CodeQL supports the following language identifiers:

LanguageIdentifierAlternatives
C/C++c-cppc, cpp
C#csharp—
Gogo—
Java/Kotlinjava-kotlinjava, kotlin
JavaScript/TypeScriptjavascript-typescriptjavascript, typescript
Pythonpython—
Rubyruby—
Rustrust—
Swiftswift—
GitHub Actionsactions—
Alternative identifiers are equivalent to the standard identifier (e.g., javascript does not exclude TypeScript analysis).

Core Workflow — GitHub Actions

Step 1: Choose Setup Type

  • Default setup — Enable from repository Settings → Advanced Security → CodeQL analysis. Best for getting started quickly. Uses none build mode for most languages.
  • Advanced setup — Create a .github/workflows/codeql.yml file for full control over triggers, build modes, query suites, and matrix strategies.

To switch from default to advanced: disable default setup first, then commit the workflow file.

Step 2: Configure Workflow Triggers

Define when scanning runs:

on:
  push:
    branches: [main, protected]
  pull_request:
    branches: [main]
  schedule:
    - cron: '30 6 * * 1'  # Weekly Monday 6:30 UTC
  • push — scans on every push to specified branches; results appear in Security tab
  • pull_request — scans PR merge commits; results appear as PR check annotations
  • schedule — periodic scans of the default branch (cron must exist on default branch)
  • merge_group — add if repository uses merge queues

To skip scans for documentation-only PRs:

on:
  pull_request:
    paths-ignore:
      - '**/*.md'
      - '**/*.txt'
paths-ignore controls whether the workflow runs, not which files are analyzed.

Step 3: Configure Permissions

Set least-privilege permissions:

permissions:
  security-events: write   # Required to upload SARIF results
  contents: read            # Required to checkout code
  actions: read             # Required for private repos using codeql-action

Step 4: Configure Language Matrix

Use a matrix strategy to analyze each language in parallel:

jobs:
  analyze:
    name: Analyze (${{ matrix.language }})
    runs-on: ubuntu-latest
    strategy:
      fail-fast: false
      matrix:
        include:
          - language: javascript-typescript
            build-mode: none
          - language: python
            build-mode: none

For compiled languages, set the appropriate build-mode:

  • none — no build required (supported for C/C++, C#, Java, Rust)
  • autobuild — automatic build detection
  • manual — custom build commands (advanced setup only)
For detailed per-language autobuild behavior and runner requirements, search references/compiled-languages.md.

Step 5: Configure CodeQL Init and Analysis

steps:
  - name: Checkout repository
    uses: actions/checkout@v4

  - name: Initialize CodeQL
    uses: github/codeql-action/init@v4
    with:
      languages: ${{ matrix.language }}
      build-mode: ${{ matrix.build-mode }}
      queries: security-extended
      dependency-caching: true

  - name: Perform CodeQL Analysis
    uses: github/codeql-action/analyze@v4
    with:
      category: "/language:${{ matrix.language }}"

Query suite options:

  • security-extended — default security queries plus additional coverage
  • security-and-quality — security plus code quality queries
  • Custom query packs via packs: input (e.g., codeql/javascript-queries:AlertSuppression.ql)

Dependency caching: Set dependency-caching: true on the init action to cache restored dependencies across runs.

Analysis category: Use category to distinguish SARIF results in monorepos (e.g., per-language, per-component).

Step 6: Monorepo Configuration

For monorepos with multiple components, use the category parameter to separate SARIF results:

category: "/language:${{ matrix.language }}/component:frontend"

To restrict analysis to specific directories, use a CodeQL configuration file (.github/codeql/codeql-config.yml):

paths:
  - apps/
  - services/
paths-ignore:
  - node_modules/
  - '**/test/**'

Reference it in the workflow:

- uses: github/codeql-action/init@v4
  with:
    config-file: .github/codeql/codeql-config.yml

Step 7: Manual Build Steps (Compiled Languages)

If autobuild fails or custom build commands are needed:

- language: c-cpp
  build-mode: manual

Then add explicit build steps between init and analyze:

- if: matrix.build-mode == 'manual'
  name: Build
  run: |
    make bootstrap
    make release

Core Workflow — CodeQL CLI

Step 1: Install the CodeQL CLI

Download the CodeQL bundle (includes CLI + precompiled queries):

# Download from https://github.com/github/codeql-action/releases
# Extract and add to PATH
export PATH="$HOME/codeql:$PATH"

# Verify installation
codeql resolve packs
codeql resolve languages
Always use the CodeQL bundle, not a standalone CLI download. The bundle ensures query compatibility and provides precompiled queries for better performance.

Step 2: Create a CodeQL Database

# Single language
codeql database create codeql-db \
  --language=javascript-typescript \
  --source-root=src

# Multiple languages (cluster mode)
codeql database create codeql-dbs \
  --db-cluster \
  --language=java,python \
  --command=./build.sh \
  --source-root=src

For compiled languages, provide the build command via --command.

Step 3: Analyze the Database

codeql database analyze codeql-db \
  javascript-code-scanning.qls \
  --format=sarif-latest \
  --sarif-category=javascript \
  --output=results.sarif

Common query suites: <language>-code-scanning.qls, <language>-security-extended.qls, <language>-security-and-quality.qls.

Step 4: Upload Results to GitHub

codeql github upload-results \
  --repository=owner/repo \
  --ref=refs/heads/main \
  --commit=<commit-sha> \
  --sarif=results.sarif

Requires GITHUB_TOKEN environment variable with security-events: write permission.

CLI Server Mode

To avoid repeated JVM initialization when running multiple commands:

codeql execute cli-server
For detailed CLI command reference, search references/cli-commands.md.

Alert Management

Severity Levels

Alerts have two severity dimensions:

  • Standard severity: Error, Warning, Note
  • Security severity: Critical, High, Medium, Low (derived from CVSS scores; takes display precedence)

Copilot Autofix

GitHub Copilot Autofix generates fix suggestions for CodeQL alerts in pull requests automatically — no Copilot subscription required. Review suggestions carefully before committing.

Alert Triage in PRs

  • Alerts appear as check annotations on changed lines
  • Check fails by default for error/critical/high severity alerts
  • Configure merge protection rulesets to customize the threshold
  • Dismiss false positives with a documented reason for audit trail
For detailed alert management guidance, search references/alert-management.md.

Custom Queries and Packs

Using Custom Query Packs

- uses: github/codeql-action/init@v4
  with:
    packs: |
      my-org/my-security-queries@1.0.0
      codeql/javascript-queries:AlertSuppression.ql

Creating Custom Query Packs

Use the CodeQL CLI to create and publish packs:

# Initialize a new pack
codeql pack init my-org/my-queries

# Install dependencies
codeql pack install

# Publish to GitHub Container Registry
codeql pack publish

CodeQL Configuration File

For advanced query and path configuration, create .github/codeql/codeql-config.yml:

paths:
  - apps/
  - services/
paths-ignore:
  - '**/test/**'
  - node_modules/
queries:
  - uses: security-extended
packs:
  javascript-typescript:
    - my-org/my-custom-queries

Code Scanning Logs

Summary Metrics

Workflow logs include key metrics:

  • Lines of code in codebase — baseline before extraction
  • Lines extracted — including external libraries and auto-generated files
  • Extraction errors/warnings — files that failed or produced warnings during extraction

Debug Logging

To enable detailed diagnostics:

  • GitHub Actions: re-run the workflow with "Enable debug logging" checked
  • CodeQL CLI: use --verbosity=progress++ and --logdir=codeql-logs

Troubleshooting

Common Issues

ProblemSolution
Workflow not triggeringVerify on: triggers match event; check paths/branches filters; ensure workflow exists on target branch
Resource not accessible errorAdd security-events: write and contents: read permissions
Autobuild failureSwitch to build-mode: manual and add explicit build commands
No source code seenVerify --source-root, build command, and language identifier
C# compiler failureCheck for /p:EmitCompilerGeneratedFiles=true conflicts with .sqlproj or legacy projects
Fewer lines scanned than expectedSwitch from none to autobuild/manual; verify build compiles all source
Kotlin in no-build modeDisable and re-enable default setup to switch to autobuild
Cache miss every runVerify dependency-caching: true on init action
Out of disk/memoryUse larger runners; reduce analysis scope via paths config; use build-mode: none
SARIF upload failsEnsure token has security-events: write; check 10 MB file size limit
SARIF results exceed limitsSplit across multiple uploads with different --sarif-category; reduce query scope
Two CodeQL workflowsDisable default setup if using advanced setup, or remove old workflow file
Slow analysisEnable dependency caching; use --threads=0; reduce query suite scope
For comprehensive troubleshooting with detailed solutions, search references/troubleshooting.md.

Hardware Requirements (Self-Hosted Runners)

Codebase SizeRAMCPU
Small (<100K LOC)8 GB+2 cores
Medium (100K–1M LOC)16 GB+4–8 cores
Large (>1M LOC)64 GB+8 cores

All sizes: SSD with ≥14 GB free disk space.

Action Versioning

Pin CodeQL actions to a specific major version:

uses: github/codeql-action/init@v4      # Recommended
uses: github/codeql-action/autobuild@v4
uses: github/codeql-action/analyze@v4

For maximum security, pin to a full commit SHA instead of a version tag.

Reference Files

For detailed documentation, load the following reference files as needed:

  • references/workflow-configuration.md — Full workflow trigger, runner, and configuration options

- Search patterns: trigger, schedule, paths-ignore, db-location, model packs, alert severity, merge protection, concurrency, config file

  • references/cli-commands.md — Complete CodeQL CLI command reference

- Search patterns: database create, database analyze, upload-results, resolve packs, cli-server, installation, CI integration

  • references/sarif-output.md — SARIF v2.1.0 object model, upload limits, and third-party support

- Search patterns: sarifLog, result, location, region, codeFlow, fingerprint, suppression, upload limits, third-party, precision, security-severity

  • references/compiled-languages.md — Build modes and autobuild behavior per language

- Search patterns: C/C++, C#, Java, Go, Rust, Swift, autobuild, build-mode, hardware, dependency caching

  • references/troubleshooting.md — Comprehensive error diagnosis and resolution

- Search patterns: no source code, out of disk, out of memory, 403, C# compiler, analysis too long, fewer lines, Kotlin, extraction errors, debug logging, SARIF upload, SARIF limits

  • references/alert-management.md — Alert severity, triage, Copilot Autofix, and dismissal

- Search patterns: severity, security severity, CVSS, Copilot Autofix, dismiss, triage, PR alerts, data flow, merge protection, REST API

平台分布

Codex

36.61%
按下载量换算3,977

Claude

28.4%
按下载量换算3,085

Cursor

16.37%
按下载量换算1,778

Gemini CLI

9.73%
按下载量换算1,057

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills